A Brief 400mm Comparison

Published December 18, 2014

Canon shooters have a bit of 400mm excitement right now. The biggest news, of course, is the release of the 100-400mm f/4.5-5.6 IS II lens, replacing the original version that’s had a very long, successful run. Not quite so much excitement was generated by the release of the 400mm f/4 DO IS II lens. It too replaces a long running lens, but one that has been considered more of a niche lens. (I’ll admit, though, it’s been one of my favorite niches. I used the 400 DO a lot over the years.)

Most surprisingly, neither lens was released at a huge price increase. The new DO II lists for $6900, compared to $6470 for the original version. The new 100-400 IS lists at $2200 compared to $1700 for the original. Both new versions promise several improvements to the older versions, but we were most interested in how good the optics might be. So, of course, we ran these through Imatest at 400mm the day they came in.

One thing I need everyone to be aware of: We tested these on the new, high-resolution film-transparency, backlit Imatest targets. These are particularly good for testing high-resolution telephoto lenses, but do give us higher resolution numbers than the larger, printed charts we use for more general testing. So, while you can compare the results of these four lenses today against each other, you can’t compare them to some older Imatest numbers for, say, the 70-200 f/2.8 lens and draw any reasonable conclusions.

We’re simply evaluating them today to see if the newer optics provide a higher resolution than the older ones. Of course, we expect them to, but it’s still worth checking. The original versions of both lenses were pretty darn good.

400mm DO IS Comparison

Canon 400mm f/4 DO IS (left) and version II (right). Roger Cicala,


We’ll separate the two types of lenses, both because it’s logical and to emphasize that since we’re testing at widest aperture, one shouldn’t directly compare the f/4 lenses to the f/5.6 lenses. Both of the DO lenses were tested at a shooting distance of 30 feet. The results were a bit shocking.  We usually don’t test single copies, but in this case I had no choice. We only had one copy of each available.

  400mm DO mtf50 400mm DO II mtf50
Center 11901490
Weighted Avg.9701350
Corner Avg. 7401100

Remember, telephoto lenses are generally much stronger in the corners than standard range lenses;  the angle of view is very narrow. So don’t compare these numbers to your 24-70mm zoom or whatever, that’s like comparing jet fighters and 747s. Different animals. And again, as I already mentioned, the test charts used for these tests are very high resolution, backlit film charts, so you can’t compare the numbers with someone else’s Imatest of a 300 f/2.8 or whatever on a printed chart.

But when just comparing these two lenses, the difference in resolution between the original 400mm DO and the Mk II version was rather amazing to me.  Canon said the new version has, “gapless dual-layered diffractive optical (DO) elements that help improve optical performance”.  Well, it certainly does.


Illustration courtesy Canon Europe.


I don’t have the knowledge to add any more explanation about how it works, but the Canon Europe Infobank  does give some more background for those who are into that kind of thing.

The difference in Canon’s computer-generated MTF charts for the 400mm DO and 400mm DO II is pretty striking, and I think our results confirm that difference shows up in actual optical testing as well as in the computer ray tracing.  Canon also says the new design has much greater contrast, and I tend to believe that too, although we’ll all want to see images in a variety of lighting conditions.

From a resolution standpoint, though, this seems absolutely a superb upgrade.  This is a night-and-day difference.

The 100-400mm Zooms at 400mm

Canon 100-400 IS L (left) and 100-400 IS L Mk II (right) fully extended.


I’d love to be presenting you a complete test of these lenses at various focal lengths, but I’ve been out sick and am still only working part time, so more complete tests will have to wait.  Because we had plenty of copies of these, though, we did test four copies of each and average the numbers. Sample variation was not very large, but it’s still just four copies. Don’t take it as gospel until we’ve tested 30 or so.

  100-400mm mtf50 100-400 II mtf50
Center 13001380
Weighted Avg.9801020
Corner Avg. 680760

The Canon computer-generated MTF charts suggest that the center of the image should be similar in both the old and new versions, but that the Mk II should be noticeably better off-axis and particularly in the corners. We do see that difference in the resolution tests, but I had expected the difference to be a bit larger. It’s certainly not nearly as dramatic as the DO difference.

From a resolutions standpoint I wouldn’t rush out to upgrade my old 100-400 IS L. The resolution difference is mostly in the corners at 400mm, and most of us are more interested in center and mid-field sharpness at that focal length.

Still, the fact that the new lens has a regular rotating zoom ring rather than a trombone barrel and much better IS system will be plenty of reason for a lot of folks to upgrade. (Personally, I liked the trombone, but I’ll give it up for the stronger IS.)

Roger Cicala and Aaron Closz

December, 2014





Author: Roger Cicala

I’m Roger and I am the founder of Hailed as one of the optic nerds here, I enjoy shooting collimated light through 30X microscope objectives in my spare time. When I do take real pictures I like using something different: a Medium format, or Pentax K1, or a Sony RX1R.

Posted in Equipment
  • novainvicta

    After using the MK1 version I finally bought the MKII and feel its much better both optically and the IS. The trombone zoom was easy on the old lens but did suck in dirt the rotary mechanism of the new lens is well weighted and can still be altered by the tighten ring. Overall in my book the MKII is better.

  • tim

    I wish the article compared the 400 5.6 and the 100-400mk ii. I’ve only seen one Youtube by Tony Northrup that really compared the image quality. He said the new 100-400 won, but could only zoom to 380mm. The 400 5.6 is so much sharper than the old 100-400mm, have thought about just getting it instead of the new lens. I rarely shoot at less than 400mm.

  • cathe roets

    I bought the new 100-400 L IS ii lens, had SERIOUS vingnetting problems on Canon 5 d iii body. Took it back where I bought the lens, they tested it and agreed about vingnetting problem. They offered me a new lens, tested it and found exactly the same vingnetting problem on second lens. I was refunded. A friend also bought new 100-400 ii lens and also had exact same vingnetting problem!!Anyone had same problem??? so disappointed !!!!

  • Jon

    I too like the push-pull of the old lens. Once you get used to it, it’s very intuitive.

  • NancyP

    Maybe. But most people like the current 400 f/5.6L for its light weight and relatively low cost. I daresay what most users would prefer would be the addition of IS, which would increase the cost and the weight a bit. A truly sharp 600 f/4 DO or 800 f/5.6 DO would be amazing for the birders.

  • Steve

    Imagine a new 400mm f5.6 DO lens. Canon would have a hard time keeping up with demand.

  • Michael

    That new DO technology seems to be really good. Nikon just updated their 300/4 with a “PF” version – it seems to be essentially same “DO” tech, but apparently like version 1 “DO” with lots of funny flare. Contrast seems ok, though. BUT it is a TINY cute little lensie.

    I wonder if Canon will follow up and design a new 300/4 DO. It would seem obvious. But what will come of the 400/5.6?

  • Mr B

    Thanks Roger, didn’t realize it was that variable! 🙁

    The 5DIII is a relatively low resolution sensor (pixel pitch), perhaps a higher resolution sensor (such as 7D2) would show more significant differences on these newer lenses?

  • Yan

    I am wondering why Canon slice the price in its top L lenses this year. It is very plausible that Canon will introduce the DO for 24-70mm f/2.8, 70-200mm f/2.8.

  • A

    As a matter of interest has the re-designed DO element improved the bokeh of the 400/4 DO II?

    Interesting about the ambient light; might it be better to include repeatable ambient light? In the real world we don’t usually shoot in the dark after-all 😉

  • Roger Cicala

    Mr B,

    There are several other things that have changed since that test (which was our first with backlit charts). We’ve found ambient light has a negative effect and have reduced it, we found that heavier support at 400mm is necessary and have beefed that up, we’re testing on 5DIII instead of 5DII.

    Which is why I always say it’s risky to compare Imatest numbers from one testing lab to another, or even within one lab over lengths of time. Computer aided target analysis absolute numbers change with any change in the lab, sometimes a lot.


  • Abimanyu

    I am getting the new 100-400L MK2, why do people whine so much at the price and don’t even think twice when spending on a new camera. It’s the lens!

  • Yan

    Focusing speed are never a problem in all L lenses. The only main different between the old and new is the IS, zoom type, and weather sealing. I prefer the old one that cost about $1300 street price/$900 used and safe $800 or even more. Within 6-12 months the price will settle down in more reasonable price $1800 or less.

  • Mr B

    Why is the center resolution for the original 100-400L shown as 1300 when in the Tamron shootout it is shown as 1000? They are both using the new high resolution backlit target so the results should be the same! Which value is correct?

  • Mr. T.

    I miss a comparison of the focusing speed on both the new and old cameras, but also between the 400mm and the 100-400mm.

    If I go from the old 100-400 to the new, focusing speed will be a large part of my rationale for doing it.

  • On my copy of the 100-400L, bought used, the zoom friction tensioner is uneven over the length of zoom, so a setting in the middle would get stiffer as you zoom out and looser as you zoom in.

    I think the rotating zoom may be the more elegant and useful solution to this problem, for me.

    It was the 100-300L that had no zoom drag setting that was by far the greater problem, as I described it, and, if I remember correctly, that lens cost about $1500 in 1990 dollars, so it was quite a bit more expensive than the 100-400L is today, in real money (about $2700 according to an inflation calculator).

  • CarVac

    If you turn the friction down on the new 100-400, can you push and pull on the hood to zoom?

  • Robert

    Thank you for publishing this information. whatever the test numbers, initial real world testing with the 100-400 II suggests the new lens produces visibly sharper images. The close focusing is amazing, and the camera (5D III) gets a focus lock much faster than before. I’m no pixel peeper, just a dude who looks at overall results. It’s no 300 2.8, but a worthwhile improvement over the old lens.

    Reading in advance about the zoom friction ring, it struck me as a little odd. Now I see it is absolutely necessary. If you carry this lens pointed down, it will extend on its own. The 24-105 does this too, but with a bigger lens it would be a real issue. The friction ring fixes that problem.

    It zooms slower than the push/pull. I was warned about this, and it’s true. If you need to zoom super fast, you may want to hang on to your old lens. The notion of the push/pull design sucking in dust is BS. The new lens extends every bit as much, just does it differently. I wouldn’t expect dust ingress problems with either version.

    Overall, I am thrilled with the 100-400 II. Can’t wait to try a 400 DO. Can’t afford one, but could be a great lens to rent.

  • Noah Wardrip

    Something folks don’t seem to be pointing out about the new 100-400 II is the amazing MFD. Going from almost 2 meters down to less than 1 meter is a game changer for me. Bryan’s results over at TDP show a bit more of an IQ difference, especially with a TC, but of course that is with many fewer samples for both the v1 and v2 lenses.

  • Jim Bracegirdle

    Frank Kolwicz mentioned his pointing the lens down or up without holding the zoom being a concern. I experienced this whenever I first used the 100-400. After a very minor adjustment to the tension ring there was no further recurrence and I have been a happy owner for quite a few years. Also the rumours from others on the various forums regarding sucking dust while zooming have never been experienced. Jim

  • pac

    I use the very good 70-300L (non DO) regularly, had the old 100-400L for test runs twice but it did not do the job. The second copy was on par with the 70-300L at 300 but again blurred at 400, so I would always prefer cropping the 300mm photo of the 70-300. Wonder if the new one will finally be worth it at 400…

  • Ale

    I wonder what kind of numbers the old 400mm f5.6L would get from the same test? I believe there are many people considering getting the new zoom if it would give the same quality. I recall reading that a good copy of the original 100-400mm was somewhat close to the prime IQ-wise, but not quite there.

  • L.P.O.

    Very interesting about the DO.

    One thought, though. I’m just wondering if the Imatest results have mostly become better just because the new lens has better contrast. The reason I’m thinking about this is that I used to own the 70-300 DO. And while fine-detail contrast on that lens seemed to lack quite a bit, the details _were_ there and could be coaxed out with a bit of a post processing. It is my understanding these are the findings from many others 70-300 DO owners, too.

    So, as said, I’m just wondering if such a dramatic difference in Imatest numbers could at least partially be due to just significantly better microcontrast, caused by the DO structure that has less scattering of light.

  • Steve

    After seeing these results from the 400mm DOii, I can see Canon expanding the DO line now.

  • Roger, contrary to your preference for the “trombone” sliding zoom on the 100-400L I find it very awkward. I’ve always found sliding zooms to be awkward, but, now that I don’t have maximum use of my arthritic fingers and thumb and have to use a wooden handle attached to the foot for support. That means that my lens-supporting hand can’t easily move with the zoomer, but at least I can grip the lens well enough to hold it for some time. The new version, if I should spring for one, would allow me to hold the grip and rotate the zoom with one finger, all at the same time.

    25 or so years ago, when I was working with a 28-80L and a 100-300L, I found the rotating zoom of 28-80 to be a hell of a lot handier, even on a tripod as both usually were. And, for situations when the lens was pointed downward, I had to invent an elastic band or foam friction ring for the 100-300 to keep the sliding zoom at my set focal length. The rotating zoom would creep a bit, I was told, but not just flop out to maximum zoom as soon as I took my hands off it for a long exposure.

    Lenses like the 100-300L and 100-400L are not consumer grade items and more thought should have gone into making the original version of the lenses useable for more than the “baby’s birthday party” level of use (hand-holding and horizontal). Pro lenses should have pro features and not require jury-rigged work-arounds, like the ones I devised, to make them suitable for the widest possible range of uses, like pointing the lens down or up without holding the zoom.

    As to the DO lens, it is now in a class that I would consider acceptably sharp for my work and comparable to my 600. Only the price is still an obstacle.

  • Deone

    Thanks, as always for your tests.
    Get well soon!

  • Jim Thomson

    Thanks Roger,always good to see your test results. Can’t wait to see how the sigma 150-600 does when you get them in.
    Merry Christmas and Happy New Year to all at Lensrentals.

  • Steve

    It is unfortunate that the tripod collar is no longer removable on the DO. The reduced weight and easier handling would have been great for handholding.

Follow on Feedly